Privacy Preserving Methods for Multisite Data Analysis without Sharing Individual-Level Data
Thursday, November 12th | Noon - 1PM
Multisite research can strengthen clinical evidence, but sharing patient-level data across institutions is often challenging. This talk introduces approaches that allow researchers to learn across institutions while keeping individual records local, including meta-analysis, evidence synthesis that accounts for population differences, and federated learning. Using clinical examples, Dr. Chang will discuss how these methods support prediction and treatment-effect research, along with practical considerations for data harmonization, privacy, and research oversight.